Duke Listens!: Visit my main blog at MusicMachinery.com

As the the long tail gets longer, filters and recommenders grow in their importance as the way people will find music.  Clearly, the most popular type of recommender right now is the social recommender. We see social recommenders all over the web. Sites like Amazon, Last.fm,  Digg, Yahoo Music all use social recommenders.

Social recommenders, however, are extremely vulnerable to  certain kinds of gaming or attacks.  At the Recommenders '06 conference last week, Dr. Bamshad Mobasher outlined two basic types of attacks: shilling - trying to promote an item and nuking - trying to demote an item.   These types of attacks are quite real. The social site Digg is under constant attack by shills trying to get their story promoted to the front page.

Bamshad pointed to an example when a loosely  organized group  who didn't like evangelist Pat Robertson managed to trick the Amazon recommender into linking his book "Six Steps to a Spiritual Life" with a book on anal sex for men.

Bamshad suggest that one  way to defend against shills and nukes is to create hybrid recommenders - recommenders that not only use social data but some inherent measure such as text or acoustic similarity.  These types of systems are typically more robust than pure social recommenders.

Bamshad highlighted some other ways that recommenders can be protected against these attacks.  As recommenders become a part of every day life for us all, making sure that the recommenders are giving honest recommendations will be increasingly important.